The release of OpenAI's GPT-3.5 and GPT-4 models, have caused quite a stir in technology circles over the past 6 months. Although a number of GPT-based products have been released into other industries (e.g., GitHub Copilot or Notion AI), the world of private equity software has been slow to catch-on, at least for now.
Here at ListAlpha, we have had a thorough opportunity to evaluate both the strengths and the weaknesses of GPT-based models. Our preliminary testing suggests that this technology can be extremely useful for investors (particularly in the areas of company screening), however needs to be used carefully due to its inherent fragility.
For the sake of this blog, we will be using Chambers & Partners (a B2B legal tech company based in UK) as a case study of a target company / private equity deal.
One of the most practical applications of GPT for PE investors is in the area of competitor analysis. In our testing, GPT has been surprisingly effective at producing robust competitive sets on target companies that also include firmographic information such as a location, size and ownership status.
In this use case the GPT model acts as both (i) a cognitive service (comparing different companies and selecting the more relevant ones), and (ii) a company reference database (producing accurate information on the identified companies).
This is obviously a useful feature for investors looking to gain a better understanding of the competitive landscape in a particular market. Additionally, because GPT is trained on the open web, it is highly effective at analyzing consumer businesses, reviews for which are referenced on blogs and forums.
In our backtesting of about 100 mid market deals, GPT significantly outperformed premium platforms such as PitchBook, Gain.pro and Crunchbase, particularly for smaller companies or niche industries with little news coverage. We will cover this topic in a seperate blog, however it is clear that this will be disruptive trend for the legacy data providers that have historically relied on large off-shore based analyst teams to produce these insights.
Another practical application of GPT for private equity investors is in the area of market and product descriptions. GPT is highly effective at producing simple and jargon-free descriptions of underlying markets and products that a company provides.
This is particularly helpful for deal makers looking to gain a better understanding of business definition, particularly when a company's website is heavy with marketing and technical jargon. By using GPT to generate clear and concise descriptions, investors can skip the tedious website browsing that is so typical to early stage screening.
In addition to competitor analysis and market/product descriptions, GPT can also be used for potential bolt-on M&A screening. By referencing the target's existing operations and market position, GPT is surprisingly good at suggesting smaller companies that could be acquired. At this stage, we were not able to get the model to suggest revenue or cost synergies, however on general M&A logic (i.e., international expansion, product complementarity) the model did well on.
This kind analysis is similar to what an M&A advisor would do typically, but with the added benefit of being able to run this earlier in the process, (at the origination / ideation stage) and being able to repeat it on multiple targets with little operating cost.
The flip-side of bolt-on M&A is the so-called "Exit Analysis" or screening for potential strategic acquirers. Similarly, GPT does well on this task, likely benefiting from the fact that larger companies have significantly more training data available on them (in the form of investor presentations and press releases).
This kind of analysis forms the basics of company screening and thus is very helpful to have the model take a stab at, even if it needs to be revised by a human later.
As you can probably guess by now, by combining all of these pieces together, GPT can be very useful in what we call "IC paper pre-drafting" - i.e., formulating the basic building blocks that can be expanded by a deal team member with more context and insight. While AI models will never fully replace investors for these tasks, we think that they will significantly accelerate the process of drafting repetitive and mundane sections of the IC report.
This trend is likely to play out in parallel to AI models accelerating all other types of writing, from emails, to marketing copies, to academic essays.
The practical applications of GPT for investors are numerous and can provide a significant competitive advantage, particularly against their peers that still do things the old fashion way. From competitor analysis to IC paper pre-drafting, GPT can help deal makers streamline their processes, make more informed investment decisions, and ultimately achieve better returns.
At ListAlpha, we have fully integrated what is currently available from OpenAI into our platform and are excited to offer a demo to interested funds. We view this as a very strategic area of investment and will be rolling our more "cognitive stack" features in the coming months, all focused on automating deal analysis and screening. Stay tuned for more.
We tailor your onboarding experience to match what you're looking for. Tell us more about yourself and your team to help us out.